diff --git a/Manifest.toml b/Manifest.toml
index 01c92d9..47df0bc 100644
--- a/Manifest.toml
+++ b/Manifest.toml
@@ -27,9 +27,9 @@ uuid = "2a0f44e3-6c83-55bd-87e4-b1978d98bd5f"
 
 [[deps.BlockFactorizations]]
 deps = ["LinearAlgebra", "Test"]
-git-tree-sha1 = "d9063bd637e50ecba604a00691434b161076459c"
+git-tree-sha1 = "0abb1500c57376d219d58492c833de3c7a949b80"
 uuid = "5c499583-5bfe-4591-9b59-c1e192d48697"
-version = "1.1.0"
+version = "1.2.0"
 
 [[deps.ChainRulesCore]]
 deps = ["Compat", "LinearAlgebra", "SparseArrays"]
@@ -192,9 +192,9 @@ uuid = "37e2e46d-f89d-539d-b4ee-838fcccc9c8e"
 
 [[deps.LogExpFunctions]]
 deps = ["ChainRulesCore", "ChangesOfVariables", "DocStringExtensions", "InverseFunctions", "IrrationalConstants", "LinearAlgebra"]
-git-tree-sha1 = "58f25e56b706f95125dcb796f39e1fb01d913a71"
+git-tree-sha1 = "a7e100b068a6cbead98b9f4e5c8b488934b7aea0"
 uuid = "2ab3a3ac-af41-5b50-aa03-7779005ae688"
-version = "0.3.10"
+version = "0.3.11"
 
 [[deps.Logging]]
 uuid = "56ddb016-857b-54e1-b83d-db4d58db5568"
diff --git a/Project.toml b/Project.toml
index 3352edf..c81c083 100644
--- a/Project.toml
+++ b/Project.toml
@@ -1,7 +1,7 @@
 name = "CovarianceFunctions"
 uuid = "b3329135-7958-41d4-bb02-e084c5a526bf"
 authors = ["sebastianament"]
-version = "0.2.2"
+version = "0.2.3"
 
 [deps]
 BlockFactorizations = "5c499583-5bfe-4591-9b59-c1e192d48697"
@@ -22,7 +22,7 @@ ToeplitzMatrices = "c751599d-da0a-543b-9d20-d0a503d91d24"
 WoodburyFactorizations = "9f1bac23-581c-4ebc-bd36-df60d764636d"
 
 [compat]
-BlockFactorizations = "1.1"
+BlockFactorizations = "1.2"
 DiffResults = "1.0"
 FillArrays = "0.12, 0.13"
 ForwardDiff = "0.10"
diff --git a/src/gramian.jl b/src/gramian.jl
index 9ba1518..10f9614 100644
--- a/src/gramian.jl
+++ b/src/gramian.jl
@@ -95,14 +95,14 @@ end
 # by default, Gramians of matrix-valued kernels are BlockFactorizations, O(1) memory complexity
 function gramian(k::MultiKernel, x::AbstractVector, y::AbstractVector, lazy::Val{true} = Val(true))
     G = Gramian(k, x, y)
-    BlockFactorization(G)
+    BlockFactorization(G, isstrided = true) # strided because every block has same size
 end
 
 # instantiates the blocks but respects structure O(n^2d) memory complexity for gradient kernel
 function gramian(k::MultiKernel, x::AbstractVector, y::AbstractVector, lazy::Val{false})
     G = Gramian(k, x, y)
     G = Matrix(G)
-    BlockFactorization(G)
+    BlockFactorization(G, isstrided = true)
 end
 
 LinearAlgebra.issymmetric(G::Gramian) = (G.x ≡ G.y) || (G.x == G.y) # pointer check is constant time